ECG - Based Emotion Detection via Parallel - Extraction of Temporal and Spatial Features Using Convolutional Neural Network

Emotion detection from an ECG signal allows the direct assessment of inner state a human. Because signals contain nerve endings autonomic nervous system that controls behavior each emotion. Besides, emotion plays vital role in daily activities human life, where we lately witnessed outbreak (COVID-19) pandemic has bad influence on affective states humans. Therefore, it become indispensable to build intelligent capable predicting and classifying emotions their early stages. Accordingly, this study, Parallel-Extraction Temporal Spatial Features using Convolutional Neural Network (PETSFCNN) is established. So, in-depth features are extracted captured suggested parallel 2-channel structure 1-dimensional CNN network 2-dimensional then combined by feature fusion technique for more dependable classification results. Grid Search Optimized-Deep (GSO-DNN) adopted higher accuracy. To verify performance proposed method, our experiment was implemented two different datasets. The maximum accuracy 97.56% 96.34% both valence arousal were gained, respectively internationally approved DREAMER dataset. While same model private dataset achieved 76.19% 80.95% respectively. results PETSFCNN-GSO-DNN compared with state-of-the-art methods. empirical findings reveal method can detect accurately better than methods potential be as affect detection.